suppose that the government introduces a new program that will cause it to spend $100 billion a year every year from now on. To pay for this, it will have to raise taxes by $100 billion a year, permanently — and if consumers take this into account, they might well cut their spending enough to offset the increase in government purchases. But suppose the government introduces a one-time, $100 billion program to repair bridges over the next year. The government will have to issue debt to pay for this, and will have to service that debt, requiring higher taxes — say, $5 billion a year. That’s a much smaller impact on consumers’ future after-tax income than the permanent program. So much less of the spending rise will be offset by a fall in consumer demand. (I’m not considering the effect of the spending in raising income, which would probably cause consumer demand to rise rather than fall.) So economic theory — Milton Friedman’s theory! — says that spending is a more effective form of stimulus than tax cuts.
This is also a very nice statement of the Barro-Ricardo equivalence proposition, which of course, is an extension of Friedman’s permanent income hypothesis.